Slashdot Mirror
The State of Solid State Storage
carlmenezes writes "Pretty much every time a faster CPU is released, there are always a few that are marveled by the rate at which CPUs get faster but loathe the sluggish rate that storage evolves. Recognizing the allure of solid state storage, especially to performance-conscious enthusiast users, Gigabyte went about creating the first affordable solid state storage device, and they called it i-RAM. Would you pay $100 for a 4GB Solid State Drive that is up to 6x faster than a WD Raptor?"
I would rather have more storage-->not more speed. I see the need for the faster processor for calculations and video editing. These industries need space and 4gigs isn't gonna cut it.
Go to the w3.org and put Slashdot.org through the validator.
I know I would!
hell, they did it a few years ago for much slower drives
Nope. I'd rather wait longer and have more capacity for less money. After all, I use Windows as my primary OS. I'm used to waiting.
Truthfully, though, if the price came down, I'd be interested in this for a Windows install, and then install all my apps and save all my docs to an external IDE.
Mercy was given to me by Christ...I must give the same to others.
...NO! :)
Game dev and music blog
Having disk in parallel will speed up your storage much cheaper. 6x faster is not significant.
The card itself goes for $150, not including any RAM. So add 4 1GB sticks of RAM and you are looking at $500+ for the whole setup. So that is about $125 per GB...ouch!
It's more like $500 considering you have to buy the RAM aswell....
I remember seeing this sort of thing way back in the DOS days. Battery backed RAM on an ISA card. Product died out because RAM was more expensive than HD.
... But I might just pay $100 for a 10 gig drive. Have my OS and other Server and Site files on their for quick access. It wasn't too long ago when an 10 gig hard drive cost $100. And one for this much speed is definately something a site admin would look at.
_Vishal www.squad9.com
4 gigs is REALLY pushing it. i think the lowest i would expect for $100 is 15-20 gigs. I think the only reason why I would spend that money is to help their sales and hope they come out with better things down the road.
Well this tech will never catch on if they can't make it affordable. Then again, it won't ever catch on if it is affordable but not worth the price.
15,000 for a 500gb solid state drive isn't affordable
100 for a 4gb solid state drive is affordable, but not worth the price.
What they need to do is make the tech better, yet affordable. What makes it so expensive to competetivly price large solid state storage devices?
On a sidenote, is anyone going to buy this drive that is 4gb and costs 100 bucks? I don't think it's much use to anyone.
$150 + (4x$90) = $510 for 4 GB of solid state storage. Definitely not worth it.
"Some days you just can't get rid of a bomb."
OH! wait.. hehe that WAS a solid-state website.. but not the kind this article talks about...
Don't anthropomorphize computers: they hate that.
I spent about $250 in 6 gigs of RAM in anticipation of this thing (RAIDing it)
Not if its called an iRam.
RamDrive, FlashDrive, etc. are all appropriate names, but iRam? Could the product name be any less descriptive?
"Would you pay $100 for a 4GB Solid State Drive that is up to 6x faster than a WD Raptor?"
Yes, yessir, I would.
Fuck No!
The performance numbers Anand came up with on this are a little disappointing, in my view. It's nice, of course, to get a few seconds quicker startup of apps or level loads, but I doubt this is really worth it to most of us at this stage (aside from the coolness factor). Once capacity of these rises enough to make them capable of replacing HDs, though, they might be really nifty in the entertainment/HTPC space due to that silent operation. Basically, an interesting concept, still not quite ready for prime time, but getting a lot closer. Worth a quick read, anyway...
4gb is enough for a distro. you can use the conventional drives for the data.
I have several machines around the office that are just fine, but have defective hard drives. This is because Dell ships the crappiest hard drives they can find (Quantum). The machines are NOT new and fast, but they run the applications that I need them to just fine. When a hard disk goes bad, I find it difficult to install a 40GB hard disk, when all I need is a couple of gigs. Some of these machines won't even support a hard drive > 30GB.
A small capacity flash drive is just what I need in this application. I would prefer that the price for a 4GB model come down a bit though. With the solid-state hard drives, these machines could last another 5-6 years!
Would you pay $100 for a 4GB Solid State Drive that is up to 6x faster than a WD Raptor?
Yes. It's big enough to let me run my OS and key programs from it... that surpasses more RAM as the biggest $100 speed boost that can be had.
No. I wouldn't pay for that. Too little GBs, too much $$. But, consider: now that it's in production by a major company, we may be on the way to practically affordable solid-state storage.
Full Article Text follows:
For years now motherboard manufacturers have been struggling to find other markets to branch out to, in an attempt to diversify themselves, preparing for inevitable consolidation in the market. Every year at Computex, we'd hear more and more about how the motherboard business was getting tougher and we'd see more and more non-motherboard products from these manufacturers. For the most part, the non-motherboard products weren't anything special. Everyone got into making servers, then multimedia products, then cases, networking, security, water cooling; the list goes on and on.
This year's Computex wasn't very different, except for one thing - when Gigabyte showed us their collection of goodies for the new year, we were actually quite interested in one of them. And after we posted about it, we found that quite a few of you all were very interested in it too. Gigabyte's i-RAM was an immediate success and it wasn't so much that the product was a success, but it was the idea that piqued everyone's interests.
Pretty much every time a faster CPU is released, we always hear from a group of users that are marveled by the rate at which CPUs get faster but loathe the sluggish rate that storage evolves. We've been stuck with hard disks for decades now, and although the thought of eventually migrating to solid state storage has always been there, it's always been so very distant. These days you can easily get a multi-gigabyte solid state drive if you're willing to spend the tens of thousands of dollars it costs to get one; prices actually vary from the low $1000s to the $100K range for solid state devices, obviously making them impractical for desktop users.
The performance benefits of solid state storage have always been tempting. With no moving parts, reliability is improved tremendously, and at the same time, random accesses are no longer limited by slow and difficult to position read/write heads. While sequential transfer rates have improved tremendously over the past 5 years thanks to ever increasing platter densities among other improvements, it is the incredibly high latency that makes random accesses very expensive from a performance standpoint for conventional hard disks. A huge reduction in random access latency and increase in peak bandwidth are clear performance advantages to solid state storage, but until now they both came at a very high price.
The other issue with solid state storage is that DRAM is volatile, meaning that as soon as power is removed from the drive all of your data would be lost. More expensive solutions get around this by using a combination of a battery backup as well as a hard disk that keeps a backup of all data written to the solid state drive, just in case the battery or main power should fail.
Recognizing the allure of solid state storage, especially to performance-conscious enthusiast users, Gigabyte went about creating the first affordable solid state storage device, and they called it i-RAM.
By utilizing conventional DDR memory modules, Gigabyte's i-RAM is a lot cheaper to implement than more conventional solid state devices. Gigabyte sells you the card, and it's up to you to populate it with memory - a definite plus for those of us who happen to have a lot of older memory laying around, especially after next year's transition to DDR2 for AMD platforms.
The backup issue is solved by the use of a battery pack that is charged by your system on the fly, although there is no disk backup available for the i-RAM.
Through some custom logic, the i-RAM works and acts just like a regular SATA hard drive. But how much of a performance increase is there for desktop users? And is the i-RAM worth its still fairly high cost of entry? We've spent the past week trying to find out...
Gigabyte sent us the f
I'd consider buying it if I were building a system that needed some fast write speed... maybe video capture. Be neato if I could get a few and stripe 'em.
I'd love to have a super quick HD for the OS because it's accessed more frequently than, say, some old data file you haven't touched in over a year.
Music, movies, documents, pictures - I don't think these need to be on solid state drives, because they're accessed just fine (except moving GB's of files still needs to be faster), but things like the OS and applications would seem to run a lot quicker if they would all be in ram-like storage.
If you use this to hold your swap and your main partition, I think the speed improvement would be well worth it! Then buy a 300GB drive for your MP3 collection and all the other junk that that doesn't need such access speed and you are set.
Switching to Linux can be an adventure!
Wow, this thing looks almost EXACTLY like the RAM add-in cards we stuck into ISA slots in the mid/late 80's for our zippy '286 and '386 based machines.
:)
Looks like they dug up an old PCB screen, added a battery backup and changed the connectors to work with modern RAM
Among other things, I handle the physical hardware design spec for my companies product (the product is software which is loaded onto a hardware to make an "appliance"). I've received emails from quite a few vendors recently offering this sort of solid-state NV storage. I think this market sector is really starting to creep forward, and these might be the kinds of "disks" we see as the norm in the not-so-distant future.
I think first off, though, these will be like caching drives - holding only the data that is most seek-time sensitive to a particular application.
-This sig intentionally left blank
I would consider using it for a 2nd drive in my G5. Perhaps, I could put one or two large apps on there and have a fast start up time for Photoshop or Final Cut Pro?
If they are able to get some more GB in to them, I would like to have one in my Powerbook to aid in power useage.
Would you pay $100 for a 4GB Solid State Drive that is up to 6x faster than a WD Raptor?
Maybe. But I can't get it. According to the article, the i-RAM costs $150 and that only gets you the card -- you still have to populate the DIMM slots. So the price of that fast solid state drive is about $400 unless you happen to have lots of spare RAM lying around unused.
If you happen to have some DDR 2200 DIMMs that can't be used in your current machine(s), then perhaps you can spend $150 to get some use out of them. Otherwise, this drive is very expensive, and not all that fast -- since it's limited to 150MBps by the SATA bus, you'd get much better performance out of the same RAM by putting it on your motherboard (assuming you can).
Note to ACs: I usually delete AC replies without reading them. If you want to talk to me, log in.
Mmmm, hyper-fast builds that don't depend on the latency of moving parts...
Dewey, what part of this looks like authorities should be involved?
Yes, I would buy one. It would make a great swap partition!
You can't re-use code, if you can't find it.
I would buy it in an instant...
But what the summary should ask: Do you want to spend 500$ for the SSD-Card plus 4*1GB Dimms... and then the answer would be a clear no (thats more than a decent budget computer in total, and i would rather put the Dimms into my motherboard than into the card (if i feel the need, i can create a ramdisk at any point later, anyway, and with 6GByte/s and 100ns , not 140MByte/s and 100us like this one)
HI O WISE PRINCE. WHT TOOK U SO DAM LONG?
That quote is straight off of the review on Anandtech. At least give credit for the quote when you are sending in a story.
can anyone say FUD?
4GB is starting to approach being large enough to hold an entire OS and all the programs a lot of users have installed. Seems like $100 to have an OS that will boot and run faster than most RAID arrays would be worth it. Nonetheless, 4GB is still only just barely big enough. Until the size at least doubles, this is only practical for a select (rich) few.
" Well this tech will never catch on if they can't make it affordable. Then again, it won't ever catch on if it is affordable but not worth the price."
"A few years ago the first Linux-based Zaurus, the SL-5500, was released for some $600 by Sharp. Today, it only costs $140 in some places online.
There must be some word for this economic principle.
---
The "are you a script" word for today is trapped.
could be useful for a triple setup, use your ram and hd as you normally would but all the crap that windows usually sticks in the vcache and swap file could be stashed on the Solid State drive. you could then feasibly dump your ram state into it when doing a shutdown and have an instant "reboot" but as the standard HD still has everything on it if the battery backup fails then you can still do a standard boot. if you use it as a speedy ramdisk too you could build a redundancy setup on your standard HD that mirrors it, (albeit not in real time, obviously) keeping your frequently accessed documents and suchlike to hand but also safe from said power failures
If you don't risk failure you don't risk success.
The logo on the box says "patent pending". Good luck. Check out DKB's Battdisk for the Amiga, from 1987 or so. http://amiga.resource.cx/exp/search.pl?product=bat tdisk
Copy Kickstart on to this, then use it to boot and you could boot an Amiga 3000 in 3-5 seconds. Wonderful device.
[Note: DKB = Dean K. Brown's company that did some real nice, and popular, hardware for the Amiga.]
-Charles
Learning HOW to think is more important than learning WHAT to think.
wouldn't this significantly (well, x6) enhance the performance of applications that require a lot of virtual memory?
That seems like it might be worth it for, say, large databases or graphics rendering.
Human being (n.): A genetically human, genetically distinct, functioning organism.
If Microsoft could make one large enought to hold XP and Office they could sell a ton to businesses. It could start/restart faster, run faster, and possiblly be less vulnerable. You would want to be able to perform flash upgrades. The best part would be that it would be very difficult to pirate. MS would love that feature. I'm sure there would still be a hdd version but this could be big if they got Dell/Compaq/Gateway to integrate it into their systems.
Ninjas don't carry tic tacs
Back in the DOS 5 and 6 days, I used to make an 8 meg ramdrive, copy the X-wing game files to that and run from there. No load times for the cut scenes or new missions, and I still had 8 meg to use for regular memory. X-wing only used 4 meg with all the options, so as long as I could get 620K free I was good to go.
There are 01 kinds of cars in the world. The General Lee, and everything else.
Before I RTFA, I would have said YES! But it looks like it uses PCI only for power; all data transfer is done over SATA-bus, which becomes the speed bottleneck at something around 150 Mbit/sec. Since that's the case, I don't see why they made it a PCI card at all... I assume the FPGA and the DDR memory require low-voltage power not offered by a normal hard-drive-style 12V molex connector. Meh.
:/
It just seems to me that the card itself is very bulky, and a similarly-priced RAMdisk with greater storage and a better form-factor is just waiting to be implemented. Oh, and it's not 4GB RAMdisk for $100, b/c you have to purchase the DDR as well
May the threads progress competently.
You will justhave to recompile...
EVERYTHING!
$100 to have 2gb set aside for ultrafast swap and 2gb set aside to load games into SS memory.
Nevermind, I see you're talking about old machines...
The thing that makes solid-state, lightning-fast storage attractive to me is digital multitrack recording. In that world, the faster your drive the more tracks can be recorded at once. The thing is, when you're recording 18 tracks of 16/44.1 PCM, 1 gig lasts about ten seconds. The same can be said about digital video. I applaud the speed, and I probably will wind up buying one, but when the capacities get as high as standard drives are now PLUS that speed, then it'll be something I won't be able to live without.
They will never stop until somebody makes the
Any clue when this will be available?
Looks like this is driven by a Xilinx 1M gate FPGA. Wonder if it could be setup to act as a coprocessor for other applications.
I would love a solid state hard-drive. But, I don't want it to be volatile like this one is. I want it to be static and I want enough read/write cycles that it doesn't fall apart after a year of heavy use.
I personally want to have my operating system on a 10GB flash drive. But, I don't want to do it with RAM because then it won't be faster to boot.
And I've got the feeling this post was an advertisement anyhow.
"The optimist proclaims that we live in the best of all possible worlds, and the pessimist fears this is true." --James
Would I be out of line to ask if you are farming your Karma?
not FUD, TROLL
Ok, so it's faster than a disk drive. How reliable is it?
*IF* if were 4gb of very reliable and very fast storage, then yes, it would be worth $100.
Mod me down with all of your hatred and your journey towards the dark side will be complete!
Because, if you had RTFPP (parent post) which you even quoted yourself, he wasn't complaining about the size but rather the price.
We already have most of that -- a ram disk. If power fails on this device, the contents are gone. The does make it a little more compatible, but why not let the system operator configure the 4gb of ram in an optimal way for their usIie -- cache v. ramdisk v. application space, etc. I have seen some ram disk/cache software that will reduce the cache space when more ram disk space is configured.
Fight Spammers!
This is incredibly sad. How many converts, exactly, are you hoping to land? This /., after all. I, for one, am still running AIX 4.3 on my home machine.
GUIs make my skin crawl.
Don't be a looter...and yes, I know that it's spelled with an "A" instead of an "E".
would power use in this drive be more or less then a conventional drive? wouldn't this be a good low power replacement for mp3 player HD's and even video polayers in the future?
GUIs make my skin crawl
Thats interesting.. Does the sun also make your skin crawl?
My partner Stuart, for instance, is addicted to capucino (has a $10,000 machine of his own at home where he does most of his work) and when he does a major kernel compile it takes his maxed-out system 5-10 minutes to complete. This is just long enough for him to go get a cup and mostly finish it off. If the compile took less than 2 minutes, he'd probably wait in his chair instead.
His recently purchased 15,000 RPM ultra 320 SCSI drives can mostly keep his current dual CPU system choked at the CPU, but if we put in one of the faster CPU quads or other system it will languish waiting for the disk.
The use of solid-state drives, even as small as 4 Gigs (although I'd probably go for 4-8 of them to increase the size and throughput by RAID0) would keep him in his chair more, and more than pay for the cost in increased productivity.
Of course if you look at this from the point of view of a server instead of a workstation, the economic reasoning may be easier.
Been there, done that, paid for the T-shirt
and didn't get it
I know which drive I'll be putting my swap file on now.
Nice troll.
It's called "lazy writes", i.e. the OS waits until all the disk buffers are full, or a time limit expires before it writes a buffer to disk. It's a pretty standard operating system optimization - Windows uses it too. "The whole sync() thing" flushes all the buffers and updates the superblock, telling the OS that the file system is "clean". Windows does this also, this is why you see CHKDSK (the Windows version of fsck) running after a rare Windows system crash.
No folly is more costly than the folly of intolerant idealism. - Winston Churchill
I use Gentoo; how does this affect me?
Uh, why wouldn't it? To the system it's just a SATA drive. I've got a gentoo system that's taking up 1.8GB and doing useful stuff. It's not very performance intensive right now, as I've only got 5 phones attached to it (it's an asterisk PBX), but should I go and attach another 100 or so, I can see how this would be pretty cool. I could store the voicemail off on some big disk, and the rest of everything would be perfectly happy on this drive.
Of course, I do wonder whether the reliability would be good. The battery in particular bothers me. On the other hand, while TFA says there's no disk-backup system, I disagree....seems like dd should work ok, right?
Given a choice between free speech and free beer, most people will take the beer.
Surely you'd be better off investing the money in more RAM for your PC, which could then use it as a cache for whatever was actually in use at the time.
My Journal
This is what I was thinking. Most of my audio and projects would fit on a 4G drive (4G would be tight for video). A drive like this would get used while I was working on the project, and then finished projects would get moved to to my online and offline storage.
(S(SKK)(SKK))(S(SKK)(SKK))
Would I buy something like this? Maybe in conjunction with a hardware encryption system, I'd have the ultimate in secure quick erase storage as well as being rugged for mobile usage.
It's a step in the right direction, but we need advances in memory size, cost, and MTBF...
If my grammar and spelling are off, I am [distracted/tired/careless] (take your pick)
The reason why this product is, er, of rather limited use is because you'll get better performance if you don't put a SATA interface between your memory and the processor for no good reason. Just put your 'Swap RAM' into the motherboard. If you don't load that many apps, the spare memory will get used for cacheing all your recently used files and you get even faster load times. Initial boot isn't as fast as all the data does have to come off the disk to start with, so all you're really gaining here is a faster boot.
OK. Some motherboards don't have space for 4GB more RAM, but if you spend $150 more than usual, they probably do.
In soviet russia stale jokes recycle you!
Would you pay $100 for a 4GB Solid State Drive that is up to 6x faster than a WD Raptor?
No : $100 is just the card, you have to provide the memory, the battery only last 16 hours, and the benchmarks are not that impressive.
The i-RAM only uses SATA for data interface... if I recall, SATA is limited to about 150 MB/sec. Raptor speed is 72 MB/sec. Where is the 6x coming from?
Other bottlenecks are sure to limit this (CPU, etc).
Until I see a way to make this actually very useful (other than having one modern game on it to get better fps), there's no way I would buy at that price.
"Trolls they were, but filled with the evil will of their master: a fell race..." -- J.R.R. Tolkien on Olog-hai
The main problem (other than the limited capacity and price) is the volatility. It has a battery pack, though if the power is out for more than 16 hours (or less, as the battery ages), it loses its entire contents. Which is somewhat precarious.
A better idea would have been to have a bank of Flash EEPROM built onto the card as a backup device, with loss of power triggering the automatic dumping of RAM contents to Flash, and resumption of power repopulating RAM from Flash on demand/during idle time. Given that it is now possible to fit 4Gb in a Compact Flash card, there is little excuse for not having such a backup subsystem.
...if that $100 would buy me a 4G flash based IDE drive instead.
I would use it to replace the 4G hard disk in my aging but faithful Libretto.
Flourescent (adj): smelling like ground wheat.
It's by no means there yet. It's still so much worse than it could actually be. For example this device probably does not support In Place Execution, a funktion avaliable to many OSes eliminating copying data around and into RAM.
A simple (software-)ramdisk might be considerably faster, and perhaps cheaper than this. And you can always boot from a clean state.
So essentially the main market for such a device are the few (32-bit) Windows-users wanting to have more speed at any cost.
1. it works with old DDR memory. You can put newer memory, but it's only gonna be clocked at 100MHz DDR (DDR-200), mostly because the SATA cable is the limiter. IOW, if you're doing an upgrade you can put your old DDR to good use.
2. Nobody says one should make it the "only" drive. It can only get you 8GB anyway. What you can do - asymmetric pairing with a regular hdd. For instance, you could have the journal of a data-ordered ext3 filesystem on this thing, and the regular part on the HDD. This way you can almost get the best of both worlds (cheap capacity *and* speed).
3. In a server configuration this thing can do miracles for throughput. Just put your DB on it, and life is gonna be so much better ...
The Raven
I would love nothing more than if my Powerbook could have four times this much storage! The $400 would pay for itself in a couple of months...
They say the PCI interface is only used for power-
Show me a version that doesn't take up valuable PCI real-estate, make it mountable in an internal 3.5" bay, and then I might be interested.
I, for one, welcome our new karma-whore sig writing overlords
From the summary it sounded like a 2.5" or 3.5" 4GB IDE drive using flash instead of an IDE emulator and battery-backed up RAM using a PCI slot for power... and no memory included!
I'd pay $100 for 4GB of flash in a PCI or hard drive form factor, for a solid-state BSD or Linux webserver.
I don't think I'd pay $100 for a 0GB hard drive emulator that takes up both a PCI slot AND an SATA cable, and I still have to populate with RAM, and that will lose all its data if you leave it off too long.
Given that you can get a 2GB Compact Flash drive for $100 or 4GB for around $200 and you can hook those up to PATA with a $40 adapter, and populating this thing to 4GB will set me back more than that... I don't see the point.
Given that this is DRAM-based and lacks a true non-volatile storage as a backing store, I wouldn't trust this with information except to use it for scratch data.
It's not as cheap as $100. If the story submitter had RTFA, the card itself costs $150, and that doesn't include the cost of equipping it with 4GB RAM, which costs around $90X4=$360. The total cost comes out to be $510.
Yes, if you can get me an MPEG2 or MPEG4 camcorder that uses it for storage and costs well under a grand.
GCHQ Quantum Insert installed. If only our tongues were made of glass, how much more careful we would be when we speak
Would you pay $100 for a 4GB Solid State Drive that is up to 6x faster than a WD Raptor?"
No, but I'd pay $100 for a 4 gig drive that is up to 6 times more reliable than a WD Raptor (or any modern hard drive with moving parts).
First off, this thing costs WAY too much in both terms of the card and terms of the memory to populate it. This board should cost about $50 not $150. I'm saying mainly $50 mostly for the fact that it comes with the lithium battery and charging feature.
Secondly, it is way too small. If it were 8GB I could use it for something like backing up dvds that play hell with hard drives and make you defrag them often. I could use this thing at that point, and so could you.
Third, for a memory based i/o board I can think of nothing more silly than to ladden it down with a disk i/o interface. It does make it more "compatible" or whatever, but it also makes this board antiquated in about two years. If they had just made the i/o controller talk directly to the cpu this board would be smoking, and probably twice the speed.
I was actually excited to read something about a product like this, but this one is not ready for prime time.
Anyone remember Boca boards? :)
- Mind
Since this should truly be a Random Access Memory device, fragmentation shouldn't be an issue for any Operation System.
Liquid State and Solidus State will be cloned shortly.
We have secretly replaced these Slashdot mods' sense of humor with a rusty nail. Let's see if they notice!!
If it is properly supported it could make a nice disk cache or swapfile. But you might just get better performance for that money from more main memory.
One possible use would be to make it faster to suspend and restore, that would be nice.
I do lots of video editing... My typical source file size is about 3 gig, and the output is generally the same.
I'm guessing a pair of these puppies would vastly improve the performance of generating transistions, filtering video, etc.
A return to the rapid boots of Atari, C64 and the like of their era.
Yes, I know they had much small OSes, but also much slower CPUs.
Windows on Cartridge. It's been a long time coming...
Yeah, I think I might have to snag a couple of these.
Dewey, what part of this looks like authorities should be involved?
So it's basically a conventional DDR SDRAM to SATA hardware RAMDISK, powered from PCI but not dependent on it and battery backed for a few hours.
At first it sounds so simple but what a brilliant idea. For years my university saved a fortune by running everything off a network boot system, using a large ramdisk as a root drive.
This sounds like the perfect hardware solution for them, that doesn't need special drivers or software configurations or even specialist hardware, just ordinary RAM chips, a cheap adaptor and a PCI slot (I reckon you could even bodge it a bit to not need any slot seeing as power is the only thing the PCI bus is used for).
Embedded stuff and home-cinema (MythTV etc.) boxes could power down their boot harddrives almost instantly and yet still keep working, storing everything on this iRAM. Then, at midnight, power up the disk to save to a more permanent storage. Or even better, constantly record to the iRAM and then re-encode at it's own pace back out to a real hard drive.
Silent, reliable as a RAM chip, the ability to replace the chips if they get faulty, the speed, the power consumption, the ease of use. This is a marvellous idea. It's obviously NOT a permanent storage device and I'd hate to have someone buying this thinking that it was, but this really deserves to take off.
I would prefer some News about the new Seagate SSDs that are expected next month.
16GB, 2,5", IDE, 55MB/s Read, 35MB/s Write, 1k$.
On a sidenote, is anyone going to buy this drive that is 4gb and costs 100 bucks? I don't think it's much use to anyone.
In the era of cheap, throwaway crap, I'm pretty much by myself when I say "I want QUALITY". So yes, I'm planning on buying several of these later today to put them in my main machines in my business. they'll be running our mission-critical cash registers.
Show me a version that doesn't take up valuable PCI real-estate, make it mountable in an internal 3.5" bay, and then I might be interested.
2 E16820160137&CMP=OTC-Froogle&ATT=Transcend+4GB+Com pact+Flash+(CF)+Card+Model+TS4GCF45
http://www.elx.com.au/item/CFIDE1
http://www.newegg.com/product/Product.asp?Item=N8
4GB, populated, probably fits in a 2.5" drive bay let alone 3.5". There's actually versions that have 2.5" connectors and mounting brackets so you can use them directly in a laptop, but I couldn't quickly google a price on them.
I RTFA and it doesn't sound like a MS Office user on Windows XP with a nice machine needs it. I think they're going after the wrong crowd!
1. I don't get why using SATA instead of PCI for data xfer is making it infinitely more useful. Dumb! Put another FPGA on there and watch scientific users grow.
2. Or put base libraries and CPAN on there, with Perl on XILINX then we're cooking! How about a benchmark with compiling a big app?
3. Obviously it would be a big win for when the network is faster than your storage. We are close to that now and some locations will have it this year (gigabit ethernet to the home). Personally I have 100 Mbps (so they say) but on my Dell Inspiron 7.5K 450MHz RH9 laptop I get 1 Mbps much more often, maybe maxed at around 10 Mbps once due to I/O and memory, plus the lack of fast sites. But if there was a service that let me download a movie at 100 Mbps or more I'd need this device. My guess is some other slashdotters already do..
4. I was thinking this might be nice when you are downloading big chunks of data for analysis, it seems you would save a lot of power, heat and time by skipping the physical movement in your hard disk.
I'd like to see some benchmarks that make use of this in cases when you would expect it to be useful. And I'd like to see it work in Linux over PCI with some more computing power on another FPGA for you to flash your own apps in.
unfortunately, this is a somewhat misplaced enthusiam. multitrack digital audio recording will indeed burn large quantities of disk, but not at the rate cited above (even adjusting for the possibility that "track" means "stereo"). 18 tracks at 2 bytes per sample at 44100 samples per second is about 1MB/sec, not 100MB/sec.
more importantly, as others have pointed out, you can massively increase disk data rates by using RAID, which provides much better bytes/sec/<monetary-unit> and is likely to for some time.
anyway, my point was really that current leading edge disk rates already provide more than enough bandwidth for all but the largest scale multitrack projects (e.g. movie soundtrack work with 60+ tracks). Track counts of 30-50 tracks are easily achievable with today's technology.
Anyone that says this isn't worth it is not very technical in my book.
An affordable 4 GB is fantastic for this kind of thing. Use your imagination:
1. Imagine how fast your system would be installed on a battery-backed up RAM drive.
2. Imagine how fast your system would be with your memory swap file installed on this.
3. Imagine how fast your database server would be with its transaction log installed on this. Hey, throw the tempdb (for SQL Server) on there as well.
4. Many other things.
If you're thinking of this as a standard hard drive to store your DivX movies and MP3 files, you're not thinking right. Solid state drives are miracles that can speed up systems beyond anything you would expect.
I'm a big tall mofo.
If you have a system without a rotating hard disk, but one of these and a LiveCD to boot from it would make a quiet system.
Would you pay $100 for a 4GB Solid State Drive
Back in 1993, I was the first person on my block to have a 1GB Fujitsu hard drive for my Sparc Classic. It cost $900! Yes, I'll pay for 4GB solid state.
an ill wind that blows no good
It isn't fair to compare the evolution of transistors to the evolution of memory devices. While the flash memory bitcell is a variation of a MOS transistor (the gate stack has a trapping layer inside it to retain charge and therefore change state), it has many different standards it has to meet. In transistor design, there are many physical occurrences (quantum mechanical tunneling, "hot" carrier generation, ...) that can hinder device operation and much effort is made to reduce these phenomena. However, memory devices often require such events to occur during normal device operation. This means that the bitcell must be able to withstand and perform well under stresses that a transistor will (ideally) never see. These issues lead to many integration and device engineering challenges that make it impractical to put solid-state memory devices on the same roadmap as traditional logic.
If this gets popular, wouldn't motherboard and chipset manufacturers start offering battery backed up ram on the motherboard? The chipset manufacturers could put control logic in for pennies and it would be up to the system builder to add a battery pack. The end user could partition some RAM for a permanent disk. Access to that disk would be at main memory speed instead of just SATA speed. The memory would cost less because you wouldn't have to have a PCI card to put it on. (Isn't it a waste to have a PCI slot used just to draw power?)
Solid.
I don't care about your karma, I don't care about what's hip. --Weird Al
I submitted this as a story back on June 4. Since it was rejected (too verbose?), I posted it to my /. journal. My main question to other folks relates to how this
would compare to using a regular ramdisk. The main deficiency with a
ramdisk is that you'd have to reload the contents every time you reboot.
Here's my article, with all its links:
Giga-byte Technology recently came out with a DRAM-based PC card that operates as a SATA hard drive. The product, iRAM, uses power from the motherboard to keep memory active when the system is shut down. During power outages, the product uses a on-board battery to retain memory for up to 90 minutes. The iRAM card is being talked about in the news (InfoWorld, itWorldCanada, engadget, PCWorld, multiplay forum) as a means of booting Windows faster. That is, you install Windows onto the iRAM drive to take advantage of the RAM's faster read-access time. Just hope that you don't lose power for more than 90 minutes.
Is boot time really that important, since many computers are on all the time? A ramdisk might have better uses, perhaps for caching frequently-accessed files such as databases and webservers. Or, if you insist on having faster bootup, instead of putting Windows on the iRAM disk, why not just store the hibernation file there?
I implemented a RAM-based database for an internet tool in 1998 to alleviate the read/write load on my local hard drive. It turned out to be a simple solution for the problem. At the time, it was just a matter of using a DOS-based ramdisk driver (ramdisk.sys). On application startup, it copied the database files to the ramdisk. During operation, everything was read/written to the ramdisk, and periodic backups were made to the physical disk. There are some inherent risks, such as loss of data during a crash since data isn't immediately written to a physical hard drive, so it may not be a great solution for a mission-critical production database. The iRAM product would make this type of database even more stable, in that the risk of loss of data is much less.
That was a while ago, so I thought I'd look into setting up a ramdisk in XP for some amusement. Follows are the results of that search. It seems that the options are relatively sparse beyond the DOS-based driver. A few freeware and commercial packages are available, though. One key factor beyond price is the size limit of ramdisk.
Microsoft's ramdisk offerings since Win2k are limited. Included with the XP OS is a ramdisk sample driver that "provides an example of a minimal driver. Neither the driver nor the sample programs are intended for use in a production environment. Rather, they are intended for educational purposes and as a skeletal version of a driver." Installation isn't simple enough for most users to benefit.
Alternatives include a shareware ramdisk, AR ramdisk (archive link: http://web.archive.org/web/20041011170408/http:/ww w.arsoft-online.de/products/product.php?id=1) (freeware, 2GB limit, discontinued, available for download here), a freeware (64MB limit) and shareware (2GB limit) version here,
Hmm, 16 hours to get power back onto the computer. Living in Florida and having a power outage during a hurricane can result in a 24 hour power loss (or longer). I think I'll pass on such a setup. Much rather have a fast solid state drive that keeps it's data.
On the contrary, I've always been amazed at the rate of price/performance evolution in HD technology.
Consider that in 1982 a 10 MB disk cost something on the order of $3500 while today you can reasonably expect to get an 80 GB disk for $50, that's a drive that has 8000x the storage for 1/70 the price or a price/MB improvement of roughly 420,000x. And, that doesn't take into account the dramatic improvement in reliability and speed (both access and interface) that the newer drives exhibit. Do you think CPUs have kept up with this?
I've heard people predict the end of moving-parts mass storage for years now, but it still seems pretty distant considering the great values we're getting with HD technology.
Point being that I don't mind waiting a little bit for my data to come back up, and from a consumer standpoint, I'm patient enough to wait a few extra seconds.
It's not that I don't see the validity of the technology, I simply don't justify the expense for myself right now.
Mercy was given to me by Christ...I must give the same to others.
What is the point in doing this if even an old PCI based stripe set can already saturate the PCI bus? PCI is about 80 Mbyte per second tops for real world hardware if nothing else has to use the bus at the same time....
http://www.hyperossystems.co.uk/07042003/products. htm#hyperosHDIIproduct
This signature is typed manually.
You idiots. This was supposed to be funny.
It all depends on the application. I remember 11+ years ago I was a developer on the NY state tax processing project. We were using Sun Sparc 20s connected to Kodak 923D scanners to scan tax returns at a very high rate (something like 72 pages per minute duplex), the barcode information obtained from the tax returns was used to move the returns into a workflow process and the images were copied from a local filesystem over to massive (at the time) fileservers. HD's at the time (even the fastest SCSI drives available) were not able to keep up with the scanner writing the barcode info and images, our custom app processing the information, and moving the images to the filesystem. Our only alternative was to use a "RAM" disk. We stuffed our Sparc 20s will all the memory they could hold (512MB) and created a 256MB filesystem in RAM. We used a thirdparty ramdisk software product for the first release which ran on SunOS 4, but Sun actually implemented a pretty slick ramdisk with Solaris 2 which we used the following year. Benchmarks at the time found that these ramdisks were some 20+x the speed of HDs of that era. I'm suprised that the Gigabyte card is only 6x faster than a HD. You'd figure it'd be more, but I guess they are using SATA as the means through which to get to the "drive", so it may be hitting some physical limitation based on the interface... I wonder if it'd be more cost effective and faster to stick an extra 4GB of memory onto the motherboard and setup a ramdisk device similar to what we had on the Sparcs... Sure it wouldn't survive reboot, but it should only be a hair slower than reading and writing from RAM itself!
For my windows systems, the OS install plus all my apps and swap space is usually no more than about 10-15gigs. The remainder of my storage is just data (mp3,mp4,spreadsheets,etc) that isn't terribly speed sensitive...well, within reason. So if I could fit my OS/apps/swap on the fast RAMdrive and then have a 400gig "slow" disk for the rest, I'd be pretty happy if it amounted to a much faster boot and faster response for the OS and applications.
Power consumption goes up when it is removed from the PCI slot, says the article. If that's so, there is a design fault somewhere - it suggests that there are floating inputs .
Panurge has posted for the last time. Thanks for the positive moderations.
PCI Express must be really hard to program for? No market? What's the deal with no one coming out with PCIe cards even though it's been out for a loooong time now? WTF? This card is a perfect candidate. A handful of ethernet card makers have slapped shit performing chipsets they yanked off $10 PCI nics or motherboards onto PCIe boards and are charging $50-100. Up from the $40 they had been charging for over six months before; bad to worse. Oh yeah guys, that'll help. Has there been an increase in desperation or suckers I missed?
Even 1x would be a big step up from craptastic SATA. 16x PCIe and at least eight DIMM slots. Then they'd have a ram drive card they could be proud of. Though prices have risen recently, 1GB are still the long running price sweet spot so they touting that their card only cost $60 and can handle 8GB, aka big jump($100+) in price from using 2GB DIMMS, makes for a 'tarded moment. I've gone on more about PCIe, but I don't know which of these two faux paus is worse.
Would you pay $100 for a 4GB Solid State Drive that is up to 6x faster than a WD Raptor?
Yeah, I would only have my OS and applications on there with everything else on a second hard drive.
FoundNews.com - get paid to blog.,
Assuming that PCI Express has the always on feature you could get much higher bandwidth by having a hardwired interface that emulates an SATA or SCSI RAID array. That would also save an SATA slot as well.
-Mind
Nope. I'd rather wait longer and have more capacity for less money. After all, I use Windows as my primary OS. I'm used to waiting.
more more more. you PC kids and your soapbox derby.
i've been running a standard system config with an 8meg SSR disk for quite a while. since the late-late 80's, even. i've gotten to the point where, for my self-built systems, silicon-based OS/App storage partition is the preferred setup, with physical disks mostly assigned to storage, not boot.
there is nothing quite so lovely as having a system where app-load is, fundamentally, a function of silicon speed, not motor speed. it can make a huge difference to the operational configuration of your machine.
of course, if your only computing experience is Windows, 8megs ain't gonna be enough.. DOS, maybe, though.
this is why LiveCD's are so lovely: once you get things working on WORM, packing it all into EEPROM is the logical next step..
; -- the corruption of government starts with its secrets. a truly free people keep no secrets. --
The purpose of SS is, yes speed. But there is another side.
Suddenly you don't have a platter spinning at 7200-15000 RPM inside your case. Plus you don't have the moving parts inside the drive to fail (good for anything more 'mission critial'). Also not being as touchy to magnets and other elements makes them great for embedded devices and specialized areas. Also, the temperature range is higher due to the lack of moving parts, making it a lot easier than spinning up a platter in the arctic.
I'm saying there are tons of benefits beyond speed. Lets be realistic- with systems coming with 1G+ of RAM, once you load up the data it's not _that_ big of a deal in most cases. Drives are fast. Anyone who's ever hibernated their system with a gig of RAM knows that it coming back up in 2-5 seconds is pretty darn good.
Can't wait for SS storage. Right now this will allow computers in the arctic and facing the elements to work more reliabily- Koisks to work with fewer failure rates, etc.
So yes- I'd buy a 4GB drive for $100 bucks given the need for it. It's not quite there for my HTPC though to keep it quiet- need about 100x more space.
PS: There is minimal benefit to having it _AND_ magnetic media, as you might as well just load it up with RAM. Still have the noise- still have the failure rates. Meh.
-M
when you see the word 'Linux', drink!
i would be willing to pay for something like this. i mean $63 a gig here http://www.micromagic.us/ta1024266.html and then that device, the speed boost would be worth it.
I see no value in this product whatsoever.
Keep your RAM where it belongs (not at the other end of a SATA cable), and use RAID if you want a faster disk...
OK, I'm not sharp enough to run the numbers so someone do it for me, please.
If I'm worried about men in black busting down my door and I have 4 gigs of liberty-hangs-in-the-balance secret data on this thing, how quickly can a disk-wiping program overwrite all the data? Since this isn't a throughput problem, it should be pretty fast, right? Given the zeal with which some TLAs are pursuing the goal of reliable forensic recovery of data from volatile media, this isn't as wacky a question as it sounds.
And another thing - I read TFA and it didn't point out if the thing could be run without the battery. For folks who want to wipe a bunch of data quickly, that should be an option. Are there any other cards out there as functional as this one (i.e. SATA, making it easy to use as a drive) yet WITHOUT battery backup?
Am I really starting to sound paranoid?
On Linux, put the memory in your motherboard and put your system on a UPS. This gives all the advantages and none of the disadvantages.
/mnt/ramfs
mount -t ramfs ramfs1
Intron: the portion of DNA which expresses nothing useful.
So we were asked "Would you pay $100 for a 4GB Solid State Drive that is up to 6x faster than a WD Raptor?"
For my swap drive, to be honest, yes. It's 6x faster and considerably cheaper than RAM.
Unfortunately, since the actual kit uses RAM, rather than use this for swap space you might as well add the RAM to your system board and forgo the need for swap space.
Some people suggest storing the OS on this. Better (again) to stick the RAM on your system board and create a RAMdisk for the OS on initial boot.
Actually, there's a thought. Why _haven't_ I done this??
$700 waaaay to much, interfacing with sata.
SSDs are actually pretty expensive piece of equipment. And for this reason, they're usually out of reach of "normal" sysadmins, i.e. "normal" datacenters, websites, high-volume and high-latency datawarehouses, offices, tech labs, etc.
;) )
I'm actually lucky to be testing one of such toys, a 16 GB, 4x2GBit Fibre Channel RamSan 320 (the one linked is a 325), made by Texas Memory Systems. (Disclaimer: neither me nor my company are related to them in any way. We're testing their equipment, evaluating a purchase.)
The price tag of this 2U rackmounted doorstopper is $40'000 (forty thousands) and change. For sixteen gigs!
128- and 256-GB units cost like small and not-too-small houses. Stackable racks up to 64 TB are sold only to governments. (Kidding, but not too much!
People is crazy to buy storage at $2000+ per GB? Well, mainly yes, but these units aren't "storage" ones; it's unwise to use them as storage area - you'd put indexes here, views at most, not tables. The real point of SSDs isn't speed, which is at times even better than RAM, but it's latency. You can access ANY byte on the grid at the same speed, more or less regardless of what the app is doing in the meanwhile.
And support for concurrent access. It's perfect when the data to be accessed must be accessed from many "data users" at once (and with "at once" I mean dozens or hundreds of concurrent apps needing dynamica data, either for reading or writing!). Having four FC boards, each one with up to two ports, each one up to 4Gbit/sec in speed, each one linkable directly to a server OR in a fabric, means you aren't in presence of an 1:1 run of the mill SAN, but an N:M beast.
When you're a cell phone provider, and must know NOW if one of your 30 million users has credit on his phone to start the call he's making, you can't wait for a drive head to find the record on disk. Also because, most likely, another million users want to know the same information at the same time.
Consider these things as "database accelerators", for whenever your latency must be really, really low and your database must be shared between lots of consumers!
The market for SSD is a very limited one. But an humongously rich one, too.
Snatching 4GB for $500 is theft, if that's a real SSD SAN-able unit.
(of course a couple of Windows versions further on we'll need them anyway, but I mean I want one right now...)
I remember those, and this looks very familiar, doesn't it? Headed for the same dustbin of history, I would bet.
I will not be buying company stock.
While you could write an app or an OS to do so, most really aren't written to take good advantage of a ramdisk. If anything is right now, I'd bet it's probably enterprise server software. The RAMdisk concept is very enticing and would seem like the fastest option, but (in Windows) it really doesn't improve most aspects of use anywhere near as well as you'd think.
....I experienced this [lack of] effect somewhat when I built my latest PC with triple 74Gb Raptors (OS, swap and programs). It does run somewhat faster in general but as far as applications go, some things speeded up noticeably, while others didn't speed up at all.
I was really bummed last time I bought a mobo that had 4 1-gig ram slots, but found that it would only run 2.5 gigs max at full speed (or something, there was some good reason not to actually put 4x1-gigs in there....). More conventional RAM for me, thanks...
Some workloads, like databases and mail systems, benefit greatly if journaling file system is used and both data and metadata are journaled to an external block device. Using battery backed memory would benefit even more.
As it happens, I'm the owner of such a setup, which is in near pristine condition. I'm missing the original Floppy that came wth the Six-Pak, and the original manual as well. If anyone want's to donate .... email me, please... sixpack at binaryblitz com thx
come to think of it, there may have been a similar setup for the Apple II, as early as 1981... Can't remember! Anyone?
Four gigs ain't bad, especially since that's almost the size of a single-sided DVD. The SATA connector means it will even work on a Mac with no special drivers. Installing a system and setting it as a boot drive could be very interesting.
It would be nice if they had a version that could take SDRAM, but it's hard to find junk SDRAM that's larger than 256K, so four slots would only be 1GB.
Still, it's not $100, even empty. Bad submitter.
--
"Open source is good." - Steve Jobs
"Open source is evil." - Microsoft
WOW, just wow. Did anyone read the "getthefacts" link. Thats the biggist load of garbage i have ever seen.
Reality is a big nasty dragon. Fortunately I don't believe in dragons.
Since this uses conventional RAM, this requires the computer to be powered up or battery power supplied to preserve the contents. If this power drains, you lose your data. Also, this product doesn't even sync contents of iRam to the hard drive for that inevitable conclusion of what should happen if the onboard battery should fail or lose power.
I will happily wait for true non-volitile solid state hard drive solutions. I don't expect hundreds of gigabytes or even terrabytes of storage, just enough to put the system kernel and regularily used applications in a much faster storage partition, which would improve system performance overall
iRam may be a novel interim solution, especially if you should have left over RAM from new computer upgrade, but to get that 4GB of storage space, you still have to spend anywhere from $500 - $800 to populate those ram slots. Combined with the $100 initial price tag, that is just too much money for something that will lose its data after long periods of disuse.
I haven't thought of anything clever to put here, but then again most of you haven't either.
if it were a SEAGATE. i don't care for WD, maxtor/quantum, who've been screwing the consumer public with anemic 1 yr warranties(i see WD just upped their warranty on enterprise class products to five years). seagate on the other hand has a 5 year warranty on all their drives(previously only 3 years for pata/sata). after seagate, hitachi/ibm, and samsung, with a 3 year warranty...
three can keep a secret, if two are dead - benjamin franklin
Market price for a 160GB drive is what, about $100. That means cost per gig is about 62 cents.
This device is $25 per gig. Too expensive. At $5 per gig I might be more willing to buy. I suspect as time goes on though we'll see the price of solid state drives come down to below the $5 per gig mark.
I bought an Amstrad NC100 "Notepad" for $20 at a flee-market a few years ago. I brought it along on my 3-month vacation to Samoa and wrote my travel diary on it. It uses a 1MB battery-backed SRAM PCMCIA card. Which means it starts instantly, and where I left off, if that's the setting I want. And I never need to think about "saving" what I write, because everything I type ends up in same place the files are. Also: no moving parts means it runs for 40 hours on 4 AA batteries. Solid state is teh w1n.
I was fairly disappointed with the anandtech review as it didn't touch only very many operations that benefit from fast random disk access. Although there are issues with RAID-0 on these right now, once they were worked out I bet you could use this to make a killer 16 GB database!
About 25 years ago, when solid state ram was beginning to come down in price, a company (I _think_ it was Amdahl) began selling 64-128 kilobyte ram drives as high-speed swap drives. This was when a machine with _lots_ of memory had perhaps one (1) megabyte of RAM. The big advantage was that they were addressed as a hard drive, not as RAM so they were outside the direct memory address space. They were selling for around a half a million dollars or so. IBM was just beginning to sell an OS for their mainframes that actually used paging. When a company was running nearly all it's computing on a mainframe, even at half a million for the RAM drive, it saved enough time to pay for itself.
Likewise, this makes sense when you may max out your physical memory, and need really fast paging.
100 bucks for 4GB's or RAM isn't back at all. No sir. TOobad the cards costs 100 bucks and 1GB or ram generally costs 100bucks a piece where I live. So would I pay 500$ for 4GB Ramdisk. THanks but no thanks.
Slashdot 1|0 Productivity
I got excited when I saw the title, and disapointed after reading the article.
This would be much cooler if:
It was the size of a harddrive and mountable like one, using normal power connectors. It would be super cool if it were the size of a 2.5" drive and used SO-DIMM, so I could use it on my laptop (Its a bit old and limited to 320MB ram, so adding even a 1 gb ram drive would help a lot as a swap drive).
It would be the coolest if instead of using DRAM it used some kind of flash memory. Something perhaps not as dense, but a lot faster. I heard some company was going to make something like that, but I havent seen it yet.
-- Senior Software Engineer, Attorney appearance services, locallawyerapp.com.
1. Solid-state disks are NOT a cure-all for any performance problems. You must have I/O bottlenecks, and more to the point, LATENCY problems. Bandwidth problems can be cured with much cheaper RAID stripes.
Let's use a metaphor of a Ferrari as a solid-state disk, and a truck as a rotating disk.
- Ferrari fast, truck slow.
- Ferarri can't carry much, truck carries loads.
- Ferrari expensive, truck cheap.
Consider a race between a Ferrari and a Truck- Who wins? Ferrari, of course. What if I add more trucks? Five trucks still won't be any faster than the Ferrari. Same with disks- A solid-state disk has no heads to seek, no platters to rotate, no latency. Even in a many-disk RAID stripe, to get to random data, a head has to seek, a disk has to spin. Period.2. Solid-state disks are more reliable than a RAM disk. You see, if you reboot a computer with a RAMdisk, the data contained on it is toast. Not so with a solid-state disk. Yes, I know, there are RAMdisk packages which periodically save the data to a non-volatile medium. Fine. But it can leave a vunerability window- If it backs up every 30 minutes, and the machine crashes on the 29th minute, you've potentially lost 29 minutes worth of data. Is that a problem? For a bank, yes. For a home user, not likely. Your mileage may vary.
3. A solid-state disk should be considered reasonably non-volatile (though I wouldn't back one up to another). A real solid-state disk will have an integrated disk drive, flash storage, battery backup, redundant power. Often times, it will have many of these features. They are reliable. They have to be, in order to justify their enterprise-class price tags.
4. Don't freak out about the price. "Oooo! The cost per gigabyte is about 10x what a rotating disk is!" So what? Nobody (intelligent) would ever advocate replacing all the rotating disks on a system with solid-state. There simply isn't a need to do so. (unless you're in zero-G, or in a tremendously high-vibration environment, etc.)
Traditionally only about 5% of one's data has bottleneck problems. Put THAT data on solid-state, and ONLY that data! A solid-state disk is a specialty tool, and should be used only where needed- Where there are latency problems. To do otherwise would be to essentially drive the above-mentioned Ferrari through a school zone. Sure it's a fast car, but other constraints (kids in crosswalk, speed limits) would keep you from realizing the performance it offers.
Heh, I can just imaging an XP installation on this. I've tried Windows XP on a small partition and I have always ended up fighting for more space, even after moving swap, My Documents and temp to another partition. Somehow it chews the space up as you install more apps. Runtime common files, Installer packages, service pack backups, System Restore snapshots etc etc. But seriously, the only place I can see solid state useful at the moment is applications requiring resistance to extreme conditions (temperature or shock) where it would compete with flash solid state drives. Comparing this product to flash solid state drives, flash drives are very slow (1.5Mbps-8Mbps) but they are also very small (2.5" and even 1.8" models. They are also more expensive than even this card + mem but they do not need any power at all to keep the data. They also have re-write limits as all flash (100k writes?). Still it is pretty clear that this product is not for industrial uses and is pretty much a market test (the article mentioning 1k only being built).
When you already have lots of RAM and your DB indexes and temp tables are constantly being swapped, this might make sense.
Information: "I want to be anthropomorphized"
I currently use a freebsd mfs based spool, with a cron job that backs it up for recovery purposes. (Yes, this leaves a bit of a window for loss. Sue me.) The performance gains, especially running a mailing list, are very noticable.
This device looks pretty tasty - it'll survive reboots and short power outages, and OS crashes, and whatnot. Even a 1 gig dedicated ram-based spool would be pretty darn good. This would be limited to SATA speeds, but.. no latency, and survivability have me seriously considering this.
Good, Fast, Cheap - Pick any two. - RFC 1925
Copy your video game to this baby.
Kill them all!
killer app.
or use for swap, temp, email directories
wake up and hold your nose
While this is great and all, it really isn't all that terribly interesting to me. What it does do for me is open the flood gates - I guarantee that if this does marginally well (or even if it tanks, who am I to say) that Asus, Epox, Abit and bunch of the other big vendors are going to jump on this bandwagon, followed up by broadcom, silicon images and other creating a dedicated chip for this instead of xilinx as well as supporting ECC memory... and of course it would be a stand-alone PCI/PCIe card instead of plugging into the existing controller on the mainboard.
Then the real fun will start as 3ware, promise, adaptec and others pick up the card n' chips and decide to do something crazy like put a SATAII or Ultra320 interface on it and a RAID chip that treats each DIMM as a channel.... while RAID 5 wouldn't be terribly interesting, RAID-0 (essentially quad-channel baby!) could put up some serious numbers. A good use for that SLI board, put one of these in that unused 8x/16x PICe slot...
Cost will be an issue, but if they could maintain the $150 price point, or even beat it, while adding all these features - this could really be some killer upcoming technology.
Ahhhhh, something to keep dreaming about.
Yes, but 16 hours is not the same as no power. This is not the same as a USB memory stick.
As long as a loss of power will cause data loss, then it is not good for anything except TEMPORARY storage. This of course gave me the idea of moving some of my spam folders to a ram device so that I may process them for identification strings.
Fight Spammers!
That way a ram drive could actually connect to a drive controller capable of making use of the thruput, plus let you put many drives on the same controller. This would be of great benefit in a *nix machine that's running a big database like Oracle. You could keep a copy of all your raw data cooked files on the regular hard drives, but run the main database engine from the copies residing on the ram drives to get much faster performance from the database. Just periodically schedule the database engine to write hot backups over to the mechanical disks plus keep all your archive logs written to the mechanical drives so that if an o/s or db engine crash happens, you'll be able to quickly recover your ramdrive instance from the mechanical drives.
Not for a kludge with an extra battery instead of true non-volatile storage. Not in a format that was geared to desktop and is not so easy to use in a wearable where solid state would really shine. Oh, I supose I might buy one as a front-end disk cache or some suc for some yet to be determined app. But not right now. Hmm. It would make a spiffy swap device.
SATA2 is not yet provided, and it's true I was wondering about getting 300MB/s from a ram module that is quite capable of that.
I thought that, maybe, the FPGA they use cannot reach such a performance yet, and it could come with next revision, when they produce their own package from end to end.
I was more wondering about some tests missing using databases.
What better test than a database, say for a small website, with few modifications to the base and the biggest problem being that hdds are a latency hell when the db is waiting for the data to be unstored....
Under linux, I know I can easily script the partioning at each reboot, and have another script syncing the db to a hdd say every 5 minutes (x% of a max 4 gigs db @10MB/s writing speed... , syncing only the last 5 minutes journal... largely possible if your are not running a Enterprise class website...)
What would be the results of this test, aka a db with almost no latency and 100MB/s bandwith ?
Wouldn't that have been more intersting than using it as a pagefile ?
It takes 40+ muscles to frown, but only four to extend your arm and bitchslap the motherfucker
This article clearly shows that this disk aren't useful for normal applications. But:
1) How on Earth could they forget to test database-performance? That is what this is made to do! If you (like I) have a less than 4GByte database-table with a lot of writes, then you need this. Otherwise more RAM seems to be the solution.
2) The bad performance with the copying a source-dir might have to do with the file-system. If every little file is stored in a big block, then you get a big overhead. Something like Reiserfs might be much better, but I can only guess as it's not tested.
There are other solid state disks with faster SCSI-interfaces and fibre channel. As the busword one put a huge "SAN"-smile on any seller and that smile goes away very fast if you ask the stupid question "Cost?", I guess those solutions are out of the question for companies with a limit on their budget.
Would you pay $100 for a 4GB Solid State Drive that is up to 6x faster than a WD Raptor?
No, but I would pay a pretty penny, though, for a durable, portable, aircraft-black-box-reliable storage device. One that will last my lifetime, take lots of damage, and never fail (well, almost never). I am amazed at how unreliable HD's are and how blase folks are about having to replace them every few years. I'd much rather have a small, slow, wickedly reliable storage device than a huge, fast, unreliable one.
Tiller's Rule: Never use a word in written form that you've only heard and never read. You will end up looking foolish.
If it was designed to go directly through the PCI or PCI-E bus, instead of having to be plugged into a PCI slot and then having it take up a whole SATA slot. It might be worth it.
That way you would get better I/O throughput. But then its techinically more challenging.
Just out of curiousity how much power do ram modules need to stay in state? Could you make a card that sits between the ram and the motherboard to effectively hibernate ur PC? So If you had a powercut you don't lose wat you are working on. A poor mans UPS and a lot less bulky.
Automation - The Car Company Tycoon Game
Back in my day we paid $2000 for 10MB and we liked it!
Actually, I'm not that old... (This was the Hyperdrive in 1984) for the Mac, but back in day people paid a lot more for a great deal less.
"I am the king of the Romans, and am superior to rules of grammar!"
-Sigismund, Holy Roman Emperor (1368-1437)
Steve, is that you?
I am not in anyway affiliated with Max Cannon
I saw an earlier review like 2 weeks ago, a link from the Inquirer to a Chinese site. I went and made a rough translation, but I guess that's for nothing.
I wonder what interesting uses people would come up with. I've heard of comments from people that I know that heavy bittorrent users can kill the HD. Anyone has experienced that?
Just as an FYI to this document... The 300MB/s SATA interface is just around the corner. I suspect that this RAM drive would benefit greatly from this new interface.
"there are always a few that are marveled by the rate at which CPUs get faster but loathe the sluggish rate that storage evolves." Not according to an article in the new issue of Scientific American. From the article Kryder's Law: "Since the introduction of the disk drive in 1956, the density of information it can record has swelled from a paltry 2,000 bits to 100 billion bits (gigabits), all crowded in the small space of a square inch. That represents a 50-million-fold increase. Not even Moore's silicon chips can boast that kind of progress."
My first thought for something like this would be to use it for squid. You'll still get a pretty good hit rate with 2 or 4GB, and even if it were to take a dump, it's not a big deal.
It'll be more useful in a few years when people have a bunch of spare 512MB+ dimms because everything has moved to ddr2.
I might like it if I had to do massive DDNS/DHCP zone transfers or I had a high performance transactional database.
This leads me though to two conclusions. Either a) all our PCs are horribly poor and cheap construction pretty much on the edge of waiting to explode if a few Gigs of solid state storage with a modicum of reliability costs as much as the whole PC
Or b) we're being sold a bill of goods on the golly geewhizbang factor and the vendor is gouging.
IMHO, the only possible use for this is to show off to your mates that your computer boots a tiny bit quicker than theirs.
I'm VERY surprised that no one has seen that this is NOT a new product. The CENATEK Rocket Drive has been around since 2001...
And it has it's own battery backup as well.
In addition, all Rocket Drives have multiple integrated power sources: an external DC input, power from the host bus or onboard UPS.
It's even been featured on ScreenSavers (TechTV) in '01. Here are it's press releases.
I think repruhsent was hoping for a Funny mod. Which, of course, he didn't get. Zealots can rarely take a joke.
With no moving parts, reliability is improved tremendously
Right, maybe you could argue that hard drives don't crash as often as the did in the past, so this is not an issue, but still I'd certainly pay a premium to have zero moving parts. Plus consider the power savings, and the resulting reduction in heat. Oh, and don't forget about the noise reduction as well.
I guess I just feel like everyone saying "it's not that much faster for the cost" is kind of missing a key point (perhaps *the* key point). It seems to me that any speed improvement is just a bonus; it's not like it's any slower than a traditional platter drive.
according to the article, this thing requires to be plugged in to a power source constantly or else the data will disappear 16 hours, unless recharged.
As a desktop user, why would I want this kind of storage? I don't want to have my data deleted just because I forgot to turn on the PC one day.
Furthermore, if this is the case, then what can this memory do that RAM can not do much better if we assume the computer is left on? this seems only as a much slower RAM with a 16 hour battery, it does not qualify as a storage unit.
All this product offers is a rechargable battery. Other than the battery, it is cheaper to buy the RAM, stick it on your motherboard and create a RAM disk with the OS. For most configurations it would probably be faster!
To fix the battery problem, buy a UPS for the price of the card.
No, especially considering that's just $100 for the card itself - you still have to populate it with the RAM, which will run you $400 using slow RAM, DDR2100, give or take whatever the deal of the week is.
This space for rent. Call 1-800-STEAK4U
We're using flash RAM for an embedded controller in extreme conditions (sounding rocket payload): we're running Linux booting off of a PCMCIA 6GB flash-RAM device. That's a little more expensive (a couple of $k for the device) but then it's a little more rugged and a lot smaller, too.
A bit flip that hoses the filesystem would put a damper on things.
This
> Would you pay $100 for a 4GB Solid State Drive that is up to 6x faster than a WD Raptor?"
No, but I'd pay $100 for a conventional FireWire 250GB external that's had to come down in price in order to compete with solid-state.
It's not a lie. It's the truth with lossy compression.
The real value in these devices is that there are no moving parts, and thus the Mean Time Between Failures is much longer. For embeded systems which live some place hard to reach (like a submarine or a satellite) a spinning hard drive can be a real liability.
The article raises the question if memory would be better in the Iram or in system memory. I say it's better in both. Windows sucks at memory allocation and hits the pagefile too often. So yes add RAm to your system TOO but it will always take a performance hit accessing the pagefile on a mechanical drive.
/. last year.
The place where I think solidstate hds are most necesary is when working with a database. Much like a suspected government system speculated on
Another nice use for the Iram would be my Photoshop and Illustrator pagefiles.
"He's a real midnight golfer"
is a solid-state drive to use up big piles of obsolete 72-pin SIMMs, hundreds at once. It could be a big rack-mount thing (size doesn't matter) with ethernet or scsi.
Alternatively just a motherboard that has really a lot of slots so it would be possible to make a really big RAMdisk. Anybody seen motherboards with more than 8 or 12 SIMM slots?
Or are there SIMM converters that permit plugging multiple SIMMs into one slot? There used to be 30-to-72-pin converters but I haven't been able to find any to combine 72-pin SIMMs.
what is $100 dollars really? 4Gb is enough to run xp and your core apps at ridiculous speeds, and yeah why not virtual memory.
what's funny is the my home communication download speed has kept up with processor speed: my 10MHz 80286 at clone had 2400 baud modem, and my 1.5 GH Xeon gets 3.5Mbit downloads...1500 times faster.
All that is needed here is a USB attachment on the I-RAM card so that I-RAM can automatically dump its contents to hard disk periodically.
A SAN or even just buying more memory for a 65 bit machine would be more effective.
I know this is offtopic, but...
-- I systematically moderate down people who describe their abuses of the mod and metamod system in their sigs. --
Doesn't that include you? Or did I miss the sarcasm?
Send email from the afterlife! Write your e-will at Dead Man's Switch.
So this is only worth it on Windows, where you spend most of your time rebooting.
Thank you. I'm here all week.
You'd be better of using ordinary RAM instead for those things, given the price is $125/GB for the new device according to this comment.
Ordinary RAM is faster and cheaper. The non-volatile solid-state storage is only useful for things that must survive loss of power.
I recently found out that, as far as solid state storage is concerned, we have been held out on. The military has actually had it since the late 80's. Of course, they paid around $10k for 10 gigs. But, it has been almost 2 decades since then. We should be purchasing a whole hell of a lot more gigs than 4 for a way less than $100! Would I still drop the cash? Not yet. But, when the gigs:price ratio gets lower, hell yes. I would love to have windows installed and running seperately on a solid state drive.
I personally like the sound of this i-RAM - it's built on a concept that I've been pondering for some time now, but I think they should package it in a standard 3.5" or 5.25" drive form-factor and use a PCI bus power adapter to bring the 3.3v over for charging. This would allow them significantly more working space for packing in RAM modules for which they are currently space limited.
But since they are in a PCI slot, I also think they made a mistake by making the interface SATA2. They should have made the thing a PCI-based ATA/SCSI compatible port. Then any O/S with support for ATA (what, like ALL of them?) would be able to detect the RAM drive through the ATA port and they would have the full bandwidth of the PCI bus at their disposal. That would be real head-turning performance. Remember: the ATA interface is not the bottleneck for ATA performance, it's the drives themselves.
I wouldn't spend the hundred bucks on the thing until they made improvements such as these to make it work closer to its potential. I think they've made a great head start, but this seems more like an R&D sample board than a commercially ready product.
It could emulate an IDE device but actually run much faster rather than dealing with that SATA bottleneck. As well as not tying up the SATA connector.
On a tangent, where have you seen AMD64 MBs that support 64GB of RAM ? I've been thinking about using something like that for getting really good performance for a relatively big DB app, but all MBs I've seen have just 4 banks, and the biggest chips are 4MB. I guess a dual proc would double that :)
According to: http://cdrinfo.com/Sections/News/Details.aspx?News ID=14213, the iRam now supports 8GB, and should have a street price of $60.
Look back up at my post, now look back down, you're on the Internet. Now look back up. I'm a signature.
I could easily use it for Gentoo's working directories for doing compilation. Or /tmp, obviously.
If this idea catches on though, Linux ought to have a new driver to automatically cache hot files on cards like this, rather than the user having to decide which part of the filesystem to mount there.
Yes of course it is faster to just have the RAM on the motherboard, but you run out of slots eventually and another level of cache doesn't hurt anything.
As I said in another thread there should also be a version that uses 72-pin SIMMs in larger quantities, because there are so many lying around unused these days.
The backup could be a large supercapacitor rather than battery, so it won't degrade over time. If they used LiIon, its useful life will probably only be about 3 years before you have to replace the battery to survive long power outages.
Actually w/ 4 GB it would be most useful to use this as a battery backed cache. For example, for compiling files. But I would much prefer to use a RAM disk for such a scheme so that I'm not limited by SATA.
When I compile, usually the HD is the bottleneck.
I'm surprised that nobody has mentioned Photoshop.
Professional retouchers often work on 500meg to 1 gig+ size files. Even with a fully populated motherboard, Photoshop will write to its disk based scratch drive. By putting the primary scratch drive on this, saving files and many other functions would be greatly speeded up.
There's been some similar products aimed at the Photoshop market, but they cost over $1,000 without ram.
A solid state disk could saturate the SATA channel several times over. I understand that it will still have low access times and that will make a big difference, but why not connect it to the motherboard in a way that allows larger bandwidth - for example through the pci or ram bus?
We've started using pc104 with compact flash for the hard drives on our blenders. Giant, truck mounted blenders that mix sand and water, or some other chemical additives to be pumped down an oil well. They get lots and lots of vibration, so solid state is the only way to go. I'd love 4GB for $100! Damn 1gb compact flash is $1000 (of you can even find them[*])
* watch, now an arse load of folks will kindly point out links to where they're available for $399 or something
Perhaps people don't understand virtual memory?
We now have 64 bit computers people.
Just buy a motherboard that can accept a lot of RAM and your shit will just get cached into ordinary DRAM which has much more bandwidth than some I/O device.
Arguments that I have seen that don't make sense:
1) Applications will startup faster
- Why? They still need to load into RAM from disk the first time. Why cripple yourself with an I/O DRAM card when you can just throw DRAM onto the high memory bandwidth memory channel on the motherboard?
2) Lower latency - Uhmm - refer to above - your shit is persistent on DISK, so it still has to load into RAM and you still need to wait.
3) Use as a write buffer. Again - use DRAM - the OS takes care of this - there's this thing called virtual memory... It was invented in the 60s or 70s...real cutting edge technology.
4) Until they start selling software on DRAM
If you have ever priced one of the EMC boxes that large databases love, $100 for 4/8 GB of solid state storage is a bargain! Use it to cache small reference tables or indexes and you will gain a fair amount of responsiveness. It's also price competitive with adding a lot of memory or upgrading the processor.
This is a good idea.
Has anyone tried this card with Linux or BSD yet? Something on an Anandtech web forum say it works just like any SATA disk (use fdisk, etc.), and the article shows some BIOS screen shots of the SATA disk, but I haven't seen anyone try it on a non-Windows operating system.
/tmp, Apache logs, small MySQL transaction-driven tables, and any other place where I want reliable fast random writes. If I could cut down the number of writes to disks, I could use slower disks (or get more use out of my current disks).
Here's is a link to some interesting test info to show how reliable the disk performs in different failure modes.
I want one for my mail spool, MySQL bin-logs,
It's an upgraded Applied Engineering RamKeeper GS!
:)
Shame on me for mentioning 1986 hardware for an Apple.
Seriously, though, I'd love to play with one of these if it could be made to run with the full 8GB. What a shame it doesn't support SATA 2.0 with its increased bandwidth.
A better solution would be the battery backed RAM disk coupled with a normal harddrive of the same size as the RAM disk. On power failure, the drive would use an embedded system to automatically backup the RAM drive to the harddrive. Of course, the battery would have to have enough power to be able to execute the backup, but that shouldn't be too hard to gauge. Hence the harddrive is used when the battery does not have enough charge to execute an emergency backup of the RAM disk.
The net result would be persistant storage that has the performance of volitile solid state storage.
as usual slashdot drones are way off. this is news worth because: 1. 4 gig is huge for solid state storage 2. the speed is inscrease at a great rate 3. MOST IMPORTANTLY - a disk that doesn't spin, WONT FUCK UP. think how much more reliable these things will be. granted 4 gig is too smal to be useful to me, and it doesn't look fast enough to be worth it. but solid state disks is there the future should lie.
If you mod me down, I will become more powerful than you can imagine....
Can't I just allocate a 4GB RAM drive on my AMD 64 Box anytime I want to have high speed storage. And by bypassing all the fake hardware layer and directly accessing the memory wouldn't this be even faster than going out to the PCI bus? Plus it's free.
Beware: SATA2 is not a standard and does not imply 3Gbs. In much the same way as USB2 does not mean 480Mbs.
SATA-IO is a basic standard with many optional parts, one of which is 3Gbs.
Look out for manufacturers claiming SATA2 (This doesn't exist, the manufacturers jumped the gun and made it up) but not 3Gbs.
BTW: the SATA overhead is about 20%, which is not included in the speed. SATA-IO at 300MBs can actually transfer at that rate, the actual signalling rate is faster @ 375MBs (?, can't remember exactly)
It's like a ramdisk, but has the advantage that it starts tiny and can grow and shrink depending on current contents to make efficient use of available RAM. It also spills over into swap space using the normal virtual memory mechanism.
My
which creates an empty
Example of how to speed up Gentoo's portage with tmpfs.
I use to run a CF card -> IDE adapter, so I can use solid state memory as my HD. The problem was my CF was about 64megs at the time, but that was enough for emBSD and/or a very small install of OpenBSD.
I ran Ipfilter at the time, and the thing was quiet and worked great. I logged to syslog on another box, and so there was virtually no I/O, otherwise the write speed on the CF was a dog.
But this, I can see as being a great firewall item.
As long as it's YOUR money, not mine.
I cried real tears when Li Mu Bai died.
you could always go with a ram-disk, gigabyte just released theirs, but i wish i could get more than 4GB on that, but it is cheaper than the SS, although volatile. What they need to do is see if they can "network" these smaller SS drives into a bigger one about the size of your standard 3.5